> On Fri, Mar 01, 2002 at 12:51:14AM -0500, Thor Lancelot Simon wrote:
> > 
> > Just to be clear, if you move the break pointer up and down, you're
> > just going to get your *own* nonzeroed memory back, not some other
> > process's.
> 
> Is this really true even if you cross VM page boundaries?
> 
(Quite offtopic from the subject, but anyway...)

Not in NetBSD; break below a page boundary frees the memory above.
The same if you munmap() memory somewhere in the break range (as our
malloc in its current existence does), it frees the memory and you 
will get back fresh memory if you access it again.

-- Ragge